RBCs progressively develop biochemical, immunologic, structural, and physiological abnormalities during storage, collectively referred to as the storage lesion. In animal models, randomization to older RBCs (42 days) results in higher mortality. However, completed randomized trials, i.e. ABLE and RECESS, essentially comparing young (median 7-day) vs middle-aged (median 28-day) RBCs, showed no difference in outcome. These data are important, since most patients are exposed to RBCs in this spectrum of age. However, they do not inform us about the safety and effectiveness of the oldest RBCs, which some patients receive. Due to pragmatic and ethical reasons it may not be feasible to conduct a clinical trial randomizing patients to the oldest blood. An NHLBI Working Group identified an unmet need for quantitative and reproducible in vivo models of RBC function, since current assessment of RBC quality is largely limited to 24 hour viability and hemolysis. Therefore, we recently completed a pilot randomized trial to explore the use of a quantitative model of strenuous exercise (VO2 max test) to measure the effects of transfusion of 2 units of autologous RBCs. We observed intriguing differences in VO2 max and exercise duration, and exercise performance was actually worse in half of the subjects who received 42 day RBCs. Our primary hypothesis is that transfused 42-day RBCs are less effective at delivering oxygen com- pared to transfuse 7-day RBCs, which is consistent with our preliminary data. Our secondary hypothesis is that 28-day RBCs are non-inferior to 7-day RBCs in this model, which may help in part to explain the RECESS trial results. We propose a cross-over design randomized trial with 3 study interventions (autologous transfusion of 2 units of 7-, 28-, and 42-day RBCs) in conjunction with VO2 max testing to test these 2 hypotheses. We will study up to 30 fit habitually exercising volunteers, who will each undergo the 3 study cycles in random order (7-, 28-, and 42-day RBC transfusion). This design will allow increased efficiency compared to a parallel group design with the same sample size because each participant will act as his/her own control. The proposed study is significant for several reasons: 1) it is likelyto confirm our compelling preliminary data showing that the oldest RBCs allowed by FDA regulations (42 days) are inferior to 7-day RBCs in a functional assay of O2 delivery. This is important since no studies have definitively shown that storage duration can affect RBC function, i.e. O2 delivery. 2) The proposed study is also significant because it may shed light on why the RECESS trial, which compared young and middle-aged RBCs, was negative. We hypothesize that 28- day RBCs are not inferior to 7-day RBCs, which are the median ages of storage for the 2 study arms in RE- CESS. 3) Finally, we suggest that this model holds promise as a rigorous and quantitative in vivo functional assay of RBC function, which can address the quality of the oldest RBCs since a clinical trial is likely not feasible and also addresses an unme need for future RBC transfusion research as identified by NHLBI.